Keyphrases Frequency Analysis From Research Articles: A Region-Based Unsupervised Novel Approach

Due to the advancement of technology and the exponential proliferation of digital sources and textual data, the extraction of high-quality keyphrases and the summarizing of content at a high standard has become increasingly difficult in current research. Extracting high-quality keyphrases and summing texts at a high level demands the use of keyphrase frequency as a feature for keyword extraction, which is becoming more popular. This article proposed a novel unsupervised keyphrase frequency analysis (KFA) technique for feature extraction of keyphrases that is corpus-independent, domain-independent, language-agnostic, and length-free documents, and can be used by supervised and unsupervised algorithms. This proposed technique has five essential phases: data acquisition; data pre-processing; statistical methodologies; curve plotting analysis; and curve fitting technique. First, the technique begins by collecting five different datasets from various sources and then feeding those datasets into the data pre-processing phase using text pre-processing techniques. The preprocessed data is then transmitted to the region-based statistical process, followed by the curve plotting phase, and finally, the curve fitting approach. Afterward, the proposed technique is tested and assessed using five (5) standard datasets. Then, the proposed technique is compared with our recommended systems to prove its efficacy, benefits, and significance. Finally, the experimental findings indicate that the proposed technique effectively analyses the keyphrase frequency from articles and delivers the keyphrase frequency of 70.63% in 1st region and 10.74% in 2nd region of the total present keyphrase frequency.